1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
|
c
c Copyright (c) 2001-2003 The Trustees of Indiana University.
c All rights reserved.
c Copyright (c) 1998-2001 University of Notre Dame.
c All rights reserved.
c Copyright (c) 1994-1998 The Ohio State University.
c All rights reserved.
c
c This file is part of the LAM/MPI software package. For license
c information, see the LICENSE file in the top level directory of the
c LAM/MPI source distribution.
c
c $HEADER$
c
c $Id: wave.f,v 6.6 2003/03/28 22:14:17 jsquyres Exp $
c
c Function: - simulates vibrating string
c - simple parallel version
c - uses MPI library for communication
c - nodes are numbered consecutively from 0
c - same program runs on all nodes
c - allows non-integral decomposition
c Reference: - Solving Problems on Concurrent Processors - Vol I
c Fox, Johnson et al
c
program wave
include 'mpif.h'
integer*4 MAXSIZE
real*8 PI
parameter (MAXSIZE = 1000)
parameter (PI = 3.14159265)
integer*4 ierror
integer*4 myid
integer*4 mypts
integer*4 nnodes
integer*4 npts
integer*4 nstep
integer*4 residual
real*8 c
real*8 delt
real*8 delx
real*8 fac
real*8 newval(0:MAXSIZE + 1)
real*8 oldval(0:MAXSIZE + 1)
real*8 pfac
real*8 values(0:MAXSIZE + 1)
real*8 x
common /valcom/ values, oldval, newval, npts, mypts
common /topo/ nnodes, myid
c
c Initialize MPI.
c
call MPI_INIT(ierror)
call MPI_COMM_RANK(MPI_COMM_WORLD, myid, ierror)
call MPI_COMM_SIZE(MPI_COMM_WORLD, nnodes, ierror)
c
c Get scalar parameters.
c
call gparam(npts, delt, nstep)
mypts = npts / nnodes
residual = MOD(npts, nnodes)
if (mypts .gt. MAXSIZE) then
print *, 'wave: Use < 1000 points on each node.'
call MPI_FINALIZE(ierror)
stop
endif
if (residual .gt. myid) then
mypts = mypts + 1
endif
c
c Set initial values along the string.
c Determine the actual starting value of k for my node.
c
fac = 2.0 * pi
k = myid * (npts / nnodes)
if (residual .gt. myid) then
k = k + myid
else
k = k + residual
endif
do 10 j = 1, mypts
x = dble(k) / dble(npts - 1)
values(j) = dsin(fac * x)
oldval(j) = values(j)
k = k + 1
10 continue
c
c I'm not sure what c and delx could be used for and I'm too
c lazy to find out.
c
c = 1.0
delx = 1.0
pfac = (c * delt / delx) ** 2
do 20 i = 1, nstep
call exchange()
do 30 j = 1, mypts
if ((j .eq. 1) .and. (myid .eq. 0)) then
newval(j) = 0.0
elseif ((j .eq. mypts) .and. (myid .eq. (nnodes - 1)))
+ then
newval(j) = 0.0
else
newval(j) = 2.0 * values(j) - oldval(j)
+ + (values(j + 1) - 2.0 * values(j) +
+ values(j - 1)) * pfac
endif
30 continue
c
c t - 1 is t, t is t + 1, ready for next cycle
c
do 40 j = 1, mypts
oldval(j) = values(j)
values(j) = newval(j)
40 continue
20 continue
c
c Output final results.
c
call outres
c
c Exit.
c
call MPI_FINALIZE(ierror)
stop
end
c
c gparam
c
c Function: - reads scalar parameters from file
c Accepts: - # simulation points (value returned)
c - time step size (value returned)
c - # simulation steps (value returned)
c
subroutine gparam(npts, delt, nstep)
integer npts
integer nstep
real*8 delt
c print *, 'Enter number of simulation points.'
c read(*, *) npts
c print *, 'Enter size of time step.'
c read(*, *) delt
c write(*, *) 'Enter number of simulation steps.'
c read(*, *) nstep
npts = 101
delt = 0.5
nstep = 100
return
end
c
c outres
c
c Function: - writes current amplitude points to file
c
subroutine outres()
parameter (MAXSIZE = 1000)
integer myid
integer mypts
integer nnodes
integer npts
real*8 values(0:MAXSIZE + 1)
real*8 oldval(0:MAXSIZE + 1)
real*8 newval(0:MAXSIZE + 1)
character filename*10
character fileno*2
common /valcom/ values, oldval, newval, npts, mypts
common /topo/ nnodes, myid
write(fileno, '(i2.2)') myid
filename = 'wave.out' // fileno
open(unit = 8, file = filename, iostat = ierrno, err = 9)
write(8, 5) (values(j), j = 1, mypts)
5 format(f8.4)
close(8)
return
9 write(6, 15) ierrno
15 format('wave: could not open file wave.out, error code = ', i12)
stop
end
c
c exchange
c
c Function: - exchanges internal boundary points with neighbors
c
subroutine exchange()
include 'mpif.h'
parameter (MAXSIZE = 1000)
integer mypts
integer npts
real*8 values(0:MAXSIZE + 1)
real*8 oldval(0:MAXSIZE + 1)
real*8 newval(0:MAXSIZE + 1)
common /valcom/ values, oldval, newval, npts, mypts
integer myid
integer nnodes
common /topo/ nnodes, myid
integer status(MPI_STATUS_SIZE)
c
c Odd nodes exchange left and then right.
c Even nodes exchange right and then left.
c
if (MOD(myid, 2) .eq. 0) then
if (myid .ne. (nnodes - 1)) then
c
c Send/receive to/from my right.
c
call MPI_SEND(values(mypts), 1,
+ MPI_DOUBLE_PRECISION, myid + 1, 0,
+ MPI_COMM_WORLD, ierror)
call MPI_RECV(values(mypts + 1), 1,
+ MPI_DOUBLE_PRECISION, myid + 1, 0,
+ MPI_COMM_WORLD, status, ierror)
endif
if (myid .ne. 0) then
c
c Receive/send from/to my left.
c
call MPI_RECV(values(0), 1,
+ MPI_DOUBLE_PRECISION, myid - 1, 0,
+ MPI_COMM_WORLD, status, ierror)
call MPI_SEND(values(1), 1,
+ MPI_DOUBLE_PRECISION, myid - 1, 0,
+ MPI_COMM_WORLD, ierror)
endif
else
if (myid .ne. 0) then
c
c Receive/send from/to my left.
c
call MPI_RECV(values(0), 1,
+ MPI_DOUBLE_PRECISION, myid - 1, 0,
+ MPI_COMM_WORLD, status, ierror)
call MPI_SEND(values(1), 1,
+ MPI_DOUBLE_PRECISION, myid - 1, 0,
+ MPI_COMM_WORLD, ierror)
endif
if (myid .ne. (nnodes - 1)) then
c
c Send/receive to/from my right.
c
call MPI_SEND(values(mypts), 1,
+ MPI_DOUBLE_PRECISION, myid + 1, 0,
+ MPI_COMM_WORLD, ierror)
call MPI_RECV(values(mypts + 1), 1,
+ MPI_DOUBLE_PRECISION, myid + 1, 0,
+ MPI_COMM_WORLD, status, ierror)
endif
endif
end
|
